This invention relates to electronic monitoring of electric motor position by utilizing a bridge amplifier circuit to measure the ratio of impedances between two motor windings, or legs of a single winding. The measurements made can control commutation or provide position and velocity feedback to a control system.
U.S. Pat. No. 3,931,553 (January/1976) to Stich et al.
U.S. Pat. No. 4,005,347 (January/1977) to Erdman
U.S. Pat. No. 4,027,212 (May/1977) to Studer
U.S. Pat. No. 4,092,572 (May/1978) to Murata
U.S. Pat. No. 4,495,450 (January/1985) to Tokizaki et al.
U.S. Pat. No. 4,654,566 (March/1987) to Erdman
U.S. Pat. No. 4,746,844 (May/1988) to MacKelvie et al.
U.S. Pat. No. 4,758,768 (July/1988) to Hendricks et al.
U.S. Pat. No. 4,882,524 (November/1989) to Lee
U.S. Pat. No. 5,191,270 (March/1993) to McCormack
U.S. Pat. No. 5,192,900 (March/1993) to Ueki
U.S. Pat. No. 5,304,902 (April/1994) to Ueki
U.S. Pat. No. 5,327,053 (July/1994) to Mann et al.
U.S. Pat. No. 5,350,987 (September/1994) to Ueki
U.S. Pat. No. 5,751,125 (May/1998) to Weiss
U.S. Pat. No. 5,821,713 (October/1998) to Holling et al.
U.S. Pat. No. 5,864,217 (January/1999) to Lyons et al.
U.S. Pat. No. 5,990,642 (November/1999) to Park
U.S. Pat. No. 6,169,354 (February/2001) to Springer et al.
U.S. Pat. No. 6,288,514 (September/2001) to Direnzo et al.
U.S. Pat. No. 6,304,045 (October/2001) to Muszynski
Conference Record of the IEEE Industry Applications Meeting (1999, p. 143), xe2x80x9cReview of Sensorless Methods for Brushless DCxe2x80x9d
IEEE Transactions on Industry Applications, Volume 28 Issue 1, (January/February 1992. p. 120), xe2x80x9cBrushless DC motor control without position and speed sensorsxe2x80x9d
IEEE Transactions on Industry Applications, Volume 30 Issue 1, (p 85, January/February 1994, p. 85), xe2x80x9cNew modulation encoding techniques for indirect rotor position sensing in switched reluctance motorsxe2x80x9d
Design News, (Apr. 8, 1991), xe2x80x9cBrushless DC Motors Yield Design Payoffsxe2x80x9d
Conference Record of the IEEE Industry Applications Meeting (1990, p. 443), xe2x80x9cAn approach to Position Sensorless Drive for Brushless DC Motorsxe2x80x9d
IEEE Transactions on Industry Applications, V37 Issue 1, xe2x80x9cEliminating Starting Hesitation for Reliable Sensorless Control of Switched Reluctance Motorsxe2x80x9d
Conference Record of the IEEE Industry Applications Meeting (1999, p 151), xe2x80x9cSensorless Brushless DC Control Using A Current Waveform Anomalyxe2x80x9d
Conference Record of the IEEE Industry Applications Meeting (1997), xe2x80x9cInitial Rotor Angle Detection of a Non-Salient Pole Permanent Magnet Synchronous Machinexe2x80x9d
Many types of electrical motors are known. All electrical motors have a stator and a moving component. In rotary motors the moving component is called a xe2x80x9crotorxe2x80x9d. In linear motors the moving component is typically called a xe2x80x9csliderxe2x80x9d. This invention applies to all electrical DC motors, including linear motors. For simplicity, the term xe2x80x9crotorxe2x80x9d is used here to refer to the moving component of all motors, and it is understood that the term xe2x80x9crotorxe2x80x9d also comprises xe2x80x9cslidersxe2x80x9d.
FIG. 1 illustrates one type of electric motor. At the center of the motor is the rotor 1 which is the moving part of the motor. The rotor contains eight permanent magnets 2 arranged as shown so that a sequence of alternating North and South magnetic poles are exposed along the outer rim. In this drawing the rotor is shown to be rotating in a counterclockwise direction.
Surrounding the rotor is the stator 3, which is stationary. The stator is made up of twelve electromagnets 4, divided up into three phases A, B, and C. All four electromagnets of phase A are driven together by the same electrical signal, and likewise for phases B and C. The apparatus to drive the three phases of electrical current is outside the motor and not shown in FIG. 1. This example motor would be termed a three-phase, eight-pole, brushless DC motor.
The principle of operation of the motor is shown in FIG. 1a, which is an expanded view of the bottom right quadrant of FIG. 1. The other three quadrants work identically to the one shown and are omitted from the drawing for clarity. At the moment shown, phase A is xe2x80x9coffxe2x80x9d (not being driven with current), thus phase A winding has no magnetic field around it. Phase B is driven with current such that a magnetic North pole is induced on the side next to the rotor. Similarly, phase C is driven in series with B to induce a South pole next to the rotor.
Two of the rotor magnets 5, 6 are shown. At this moment, rotor magnet 5 is positioned between the phase B electromagnet and the phase C electromagnet. The South pole on the phase C electromagnet attracts the North pole on magnet 5, while the North pole on the phase B electromagnet repels North pole on magnet 5. These magnetic forces both act to push magnet 5 to the right, thereby imparting counterclockwise torque to the rotor. This force makes the motor rotate. At this time, rotor magnet 6 is positioned near the phase A electromagnet that is switched off, so there is little or no magnetic force on magnet 6.
As the rotor moves counterclockwise, the magnet 5 moves to the right until it reaches a position adjacent to electromagnet C, where the magnetic forces from windings B and C are no longer effective to push it along. To maintain motor torque at this rotor position, the pattern of currents through the phase electromagnets must be changed. Phase C will then be switched off, and phase A will be driven such that it generates a South pole next to the rotor. At this time, rotor magnet 6 is positioned between electromagnet A and B so magnet 6 is subject to forces pushing it to the right, similarly to the situation with magnet 5 earlier. Since magnet 5 is now next to phase electromagnet C and phase C is switched off, there will be no effective force on magnet 5 in this position.
As the rotor moves still further, a new rotor magnet will move into this quadrant, this one with a North pole exposed at the rim. Following that, another magnet with a South pole will move in. After 90 degrees of rotation, the situation will again appear the same as it did at the start, with a South pole on the rotor adjacent to phase winding A. During this 90 degree rotation, the phase currents will have been switched six times to keep the magnetic field applied appropriately for the rotor magnet positions. This sequence of six states is termed xe2x80x9c360 electrical degreesxe2x80x9d and is repeated for each pole pair, so we see that for an eight-pole motor, 360 electrical degrees correspond to 90 mechanical degrees of rotation. The process of switching phases to correspond to rotor position is called xe2x80x9ccommutationxe2x80x9d.
FIG. 1b illustrates the commutation sequence for the above example, showing the correspondence of rotor position to phase winding state. Phase A begins in the xe2x80x9coffxe2x80x9d position and phase B is not being driven, as described previously. The sequence of driving phase A, B, C is depicted showing the magnetic poles induced (North or South). The movement of the rotor is also shown relative to the phase A starting position.
The commutation process described above is a simple switching process sometimes called xe2x80x9ctrapezoidal drivexe2x80x9d of a motor. It is sometimes desirable to use a more complex driving method where the motor windings are driven with an arbitrary analog waveform such as a sinusoid, rather than simply being switched on and off. This is often termed xe2x80x9csinusoidal drivexe2x80x9d.
The example motor in FIGS. 1, 1a represents one common configuration for a rotary motor. Many other configurations are in common use. For example, the rotor magnets may be made up of electromagnets instead of permanent magnets; in this case the stator may or may not use permanent magnets. Other possible configurations include having the stationary stator inside the rotor or alongside it in the axial direction. The disclosed invention can be applied in all these configurations.
FIG. 1c illustrates an example of a linear motor. It has a three-phase stator 8 made up of windings, which surround the slider 7. Operation of this linear motor is very similar to that of the rotary motor just described. The three stator phases A, B, and C correspond to the phases in the rotary motor""s stator, and the permanent magnet poles in the slider correspond to the rotor magnets. The number of stator windings and rotor poles is arbitrary, and depends on the needed physical length of travel.
As with the rotary motor, linear motors can have many other configurations, and the disclosed invention is applicable to all.
The examples illustrated above (FIGS. 1, 1a, 1b, 1c) describe one method of commutation. There are different schemes for driving electric motors, but they all share the common concept of controlling the electromagnet drive currents to generate a moving magnetic field that is synchronized to the rotor position. Accordingly, the mechanical position of the rotor relative to the stator must be known by the driving apparatus in order to provide proper control of the windings. Many types of position sensing apparatus have been used with electric motors
The simplest form of position sensing in DC motors is the brush commutator. The brush commutator is still extensively used but suffers from some disadvantages as a result of friction and wear between the brushes and the commutator surfaces, and consequential reliability and maintenance problems are the result.
Brushless DC motors avoid these problems by performing commutation electronically. Electronic commutation has traditionally required the use of external position sensors mounted on the motor. The most common of these sensors are Hall-effect magnetic sensors mounted near the rotor. This technique is described in many places (i.e. ref. U.S. Pat. Nos. 4,092,572 and 4,758,768). If greater angular resolution and accuracy is needed, an optical shaft encoder is sometimes used in addition to, or instead of, Hall-effect sensors. Use of an optical encoder for commutation is described in ref. U.S. Pat. Nos. 4,005,347 and 4,882,524. Another standard sensor technology is a magnetic resolver. Other types of external sensors have also been used or suggested. ref. U.S. Pat. No. 3,931,553 describes the use of a capacitative rotation sensor for commutation control; ref. U.S. Pat. No. 5,864,217 describes use of a toothed wheel and magnetic pickup sensor; and ref. U.S. Pat. No. 4,027,212 describes techniques for motor commutation controlled by external rotation sensors in general.
Motor position sensing can also be done with extra windings built into the motor instead of using external sensors. Several approaches have been suggested; one example is described in ref. U.S. Pat. No. 6,169,354.
All of the above methods add extra cost to the system and take up extra space in or near the motor. In addition, some of the aforementioned methods have accuracy and reliability issues. To avoid these liabilities, many ideas have been previously pursued in order to find ways of eliminating extra position sensing components.
The most common approach for sensorless control of rotary motors is to sense the motor rotation by monitoring of the induced voltage in the motor windings caused by the rotating magnetic field of the rotor. This voltage waveform (termed back-EMF) is usually monitored in a motor winding leg that is not being driven; the winding used for voltage waveform monitoring shifts as the motor commutation rotates between the windings.
FIG. 2 depicts three motor windings 10, 11, 12 of a typical 3-phase motor. Windings 10, 11, 12 are connected at center node 13. A current through two of the three windings drives the motor. At any given time, two of the windings are driven in this way and the third remains idle. The example depicted in FIG. 2 shows current being sent through windings 10, 12 in the direction of arrows A and B. The third winding 11 is idle (or un-driven). Differential amplifier 20 measures the induced voltage across idle winding 11 and generates output signal 21. Output signal 21 is thus a measurement of the induced voltage across idle winding 11, and is dependent on both the speed and position of the rotor. As the motor moves, commutation causes the drive windings to be switched, and the sense winding must also be switched correspondingly.
A major disadvantage is that this method works only when the rotor is rotating at a reasonable speed, since there is no induced voltage from a stationary magnetic field. Some special technique must thus be used to get rotation started. This is acceptable in some applications such as fans and disk drives that use a constant motor rotational speed when operating, but it is unacceptable for many other applications such as robotics and tape drive applications where the motor must remain under close control when being held in a stationary position. Back-EMF sensing is commonly employed in many applications where these drawbacks are acceptable. Variations on this concept are well known in the present art. They are described in many places including the reference U.S. Pat. Nos. 6,304,045, 4,495,450, 4,654,566, and 4,746,844, and also in many published articles.
Several methods of rotor position-sensing have also been suggested that involve adding position-sensing windings to a motor (ref. U.S. Pat. No. 6,169,354). However these methods also add undesirable cost and complexity to the motor.
Some research and experimentation has been done with other sensorless motor drive techniques that use measurement of impedance variations in the windings to derive the motor mechanical position. These impedance changes take place as the magnetic poles of the rotor pass by the poles of the stator. Generally these methods involve inducing a test signal into the motor in addition to the actual motor drive currents, and measuring the high-frequency response. These signals may be pulses, as disclosed in. ref. U.S. Pat. No. 6,288,514, or the signals may be continuous AC signals, as discussed in ref. U.S. Pat. No. 5,990,642.
Several other proposed sensorless methods use dynamic measurements of winding current and applied voltage to derive mechanical position of the motor. The motor commutation is driven based on an estimated/extrapolated motor position, and the measured voltage and current parameters are used to correct the estimate through one of a variety of mathematical techniques including fundamental machine equations, dynamic models, and xe2x80x9cobserversxe2x80x9d. An xe2x80x9cobserverxe2x80x9d in this context could also be called a xe2x80x9cstate observerxe2x80x9d, and refers to specific mathematical technique(s) that consist of a mechanism (usually implemented in software) that monitors parameters of the system in operation (i.e. motor and motor-controller) and derives information that can""t be directly measured.
Ref. U.S. Pat. No. 5,751,125 describes a more novel position-sensing scheme as part of an artificial heart mechanism. It derives motor position from measurements of the inductance ratio between adjacent windings in a delta-connected motor. The winding inductances vary with motor rotation. As one winding of the delta circuit is driven, the other two un-driven windings form a voltage divider, and the voltage of the un-driven node changes with the variations in the winding inductances in that voltage divider. In this ref. patent, the position sensing is used only to control commutation; no velocity or high-resolution position measurements are needed for this application.
The ref. U.S. Pat. Nos. 5,304,902 and 5,192,900 describe a circuit configuration somewhat similar to the aforementioned artificial heart mechanism, controlling commutation based on voltage measurements at an un-driven leg or at the center node of a Wye-connected motor. However both of these patents focus on the voltages from back-EMF rather than from inductance variations.
Many papers and articles have been published exploring different methods for sensorless motor control. A useful summary was presented at the 1999 IEEE Industry Applications Meeting (1999), titled xe2x80x9cReview of Sensorless Methods for Brushless DCxe2x80x9d.
What is needed is a sensing mechanism for motor position control which does not require external sensors, and which reduces cost and improves reliability. The present invention addresses these needs and provides for measurement verification using standard methods which are known.
The main aspect of the present invention is to provide a motor position sensing mechanism for DC motors that does not require external sensors to be attached to the DC motor.
Another aspect of the present invention is to provide for a motor position sensing mechanism that does not inject any extra signals or currents into the motor mechanism.
Another aspect of the present invention is to provide good speed and positional feedback to the controlling circuitry.
Another aspect of the present invention is to insure good position control for the motor mechanism, even when the motor is stopped, idle or actively maintaining its position via a controlling servo.
Another aspect of the present invention is to insure that the motor positional control functions under any condition from stalled to unloaded.
Another aspect of the present invention is to employ a bridge amplifier circuit to measure the ratio of impedance between the two motor windings as these windings move in relation to a fixed magnet.
Another aspect of the present invention is to employ sensing means functioning to measure and verify a motor position and/or speed measurement sensed by the present invention.
Other aspects of this invention will appear from the following description and appended claims, reference being made to the accompanying drawings forming a part of this specification wherein like reference characters designate corresponding parts in the several views.
The present invention employs a bridge amplifier circuit to measure the ratio of impedances between two motor windings, or two legs of a single winding. The present invention utilizes dynamic changes in that impedance measurement to accurately track the position of the motor.
The present invention provides a circuit and method for position sensing of DC motors without requiring external sensors to be attached to the motor. The present invention does not inject extraneous signals or currents into the motor for position detection. Position sensing must be accurate enough to provide good control for motor commutation, and must provide good speed and position feedback for a servo controller. Furthermore, position sensing must work even when the motor is stopped in either an idle mode or actively maintaining its position via a position servo. Position sensing of the present invention must work under any motor-loading condition from stalled to unloaded and free-running.
The present invention employs the use of a novel circuit to perform a winding impedance measurement that avoids many of the aforementioned problems that were discussed concerning the earlier sensorless schemes.
The key feature of the present invention is that it uses a bridge amplifier circuit to measure the ratio of impedances between two motor windings, or two legs of a single winding, rather than attempt to make an accurate absolute impedance measurement.
U.S. Pat. No. 5,751,125 measures the inductance ratio between two windings, but it uses an absolute voltage measurement instead of a bridge amplifier. An absolute voltage measurement or a measurement against a fixed reference voltage will be very sensitive to a number of secondary effects that will cause serious measurement errors. Changes in temperature, fluctuations of the supply voltage, and changes in other parameters influence the measurement. Bridge amplifiers using a reference network are used to measure impedances in some sensors such as strain gauges, to compensate for similar errors. Bridge amplifier techniques can also be applied to the motor winding impedance-sensing problem. The use of such conventional methods in conjunction with the winding impedance measurement of the present invention provides means for verifing the motor positioning sensed by the present invention.